Quite the freighted term, “evolution,” especially when used by National Public Radio. I somehow doubt NPR used the word in any other context.
You are taken at face value an ICR article that's trying to decieve you.
For where the evolution part is relevant
Belcher (and Chiang) uses a method sometimes called "directed evolution," which allows her to quickly modify viruses to work with a range of materials.
In this case, directed evolution begins with a small vial that Chiang pulls from a refrigerator. Inside is a clear fluid that contains a billion viruses; they are nearly identical, but each has a subtle genetic variation introduced by the researchers. The variations are, in part, fortuitous: the researchers add a randomly generated sequence of DNA to each virus. But the added DNA, which codes for a short strand of amino acids called a peptide, is inserted into the gene for a select protein.
Since there are so many variations among the viruses in the vial, some of them should randomly have peptides that bind to a useful inorganic material. The researchers simply pour the contents of the vial onto a target material, such as a small square of gold, and give the viruses a chance to bind. Then they wash the material. After a few repetitions, only the viruses that happen to bind strongly remain. The process allows the researchers to quickly engineer viruses to bind to a particular material, even if they don't know ahead of time what sequence of amino acids is likely to work.
Funny how the ICR deciced to comment on a science article by a reporter who probably flunked JR high school biology instead of using more detailed technical reports.